Skip to content

Propagate simulator failures from make run - #42

Open
BerkmKuzu1 wants to merge 1 commit into
pulp-platform:mainfrom
BerkmKuzu1:fix/simulator-exit-status
Open

BerkmKuzu1 wants to merge 1 commit into
pulp-platform:mainfrom
BerkmKuzu1:fix/simulator-exit-status

Conversation

@BerkmKuzu1

Copy link
Copy Markdown

The batch run recipe can report success when vsim fails. Its commands are separated by semicolons, so the final if grep ... returns 0 when the transcript lacks errors happened, masking the simulator's exit status. Both CI and the regression runner use this target.

Chain cd, vsim, and the existing transcript check with &&. A failed directory change or simulator invocation now makes the target fail; successful runs still reach the transcript check.

Reproduction with Questa on PATH, from the repository root:

scratch=$(mktemp -d)
make run gui=0 QUESTA= BUILD_DIR="$scratch" VSIM_PARAMS=-codex_invalid_option
echo $?

Questa 10.7c rejects the option with exit status 1. Before this change, make returns 0; afterward it returns 2 and reports Error 1.

Validation with GNU Make 4.3 / Bash on WSL and native Questa 10.7c:

  • Controlled simulator exits 1/17, a missing executable, and a missing build directory: previously passed, now fail. The missing-directory case no longer invokes the simulator.
  • Successful exit with a benign transcript still passes; exit 0 with errors happened still fails.
  • Real Questa invalid-option reproduction confirmed before/after; a standalone $finish(0) smoke test confirmed a successful Questa exit.

These checks cover command/error handling. The full Neureka RTL regression was not run.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant